230 M. Deville on Diffusion through Iron Tubes. 



diameter. To the ends of this tube, which was thus constructed 

 without soldering", two thinner copper tubes were soldered, and 

 in these glass tubes were cemented. The whole tube was placed 

 in a porcelain tube in a furnace, and one of the ends connected 

 with a hydrogen-apparatus, while the other was provided with a 

 long vertical delivery-tube which dipped under mercury. While 

 the tube was kept at a very high temperature, hydrogen was 

 passed through until its action on the sides of the tube must 

 have been terminated and all atmospheric air and moisture com- 

 pletely expelled. When now the hydrogen-tube was sealed, the 

 mercury rapidly ascended in the vertical tube at the rate of 3 to 

 4 centimetres in a minute, and more rapidly according as the 

 heat of the furnace was increased. Hence in the interior of the 

 apparatus an almost vacuous space was formed in consequence 

 of hydrogen passing through the tube in opposition to the ex- 

 ternal atmospheric pressure. 



Deville has continued* on these thick tubes of iron the expe- 

 riments which, in conjunction with Troostf, he had previously 

 made on porous earthen tubes, and has arrived at some unex- 

 pected results. 



To the two ends of an iron tube about 3 millims. in thickness 

 were soldered two very fine copper tubes, by which it communi- 

 cated on the one hand with a source of nitrogen, and on the 

 other with a manometer. Two good stopcocks (immersed for 

 safety's sake in cold water) were cemented to the ends of these 

 copper tubes ; by one of them a current of nitrogen could be 

 introduced or stopped at will, and by the other, which was a 

 three-way stopcock, the interior of the tube could be connected 

 either with a manometer or with a mercury or a water pneu- 

 matic trough to collect the gases and analyze them. 



The iron tube was introduced into an impermeable porcelain 

 tube very slightly longer than itself. This was closed at each 

 end by a cork perforated so as to permit the copper tube to pass, 

 and at each end a glass tube was fitted so as to allow a current 

 of any gas to enter the annular space between the porcelain 

 and the iron tube. The middle of this apparatus was fastened 

 firmly in a furnace fed by gas-coke, and by a ventilator which 

 renders the operator entirely master of variations in the tempe- 

 rature. 



In this manner there could be introduced into the iron tube, 

 and into the annular space which surrounds it, two separate 

 currents of gas isolated by a metallic disphragm several millime- 

 tres in thickness. 



At first pure nitrogen was passed into the iron tube and into 



* Comptes Rendus, July 18, 1864. t Phil. Mag. vol. xxii. p. 61. 



